Design of a control strategy for guiding collaborative robots using wearable EMG and IMU sensors

Abstract

The use of robots is widespread in many industrial applications, but they can be considered as a risk if collaboration with a worker is required. In the last years, researches and studies over human-robot interaction (HRI) have increased and a new generation of robots has been developed: collaborative robots. Collaborative robots are born to work safely with humans in the same working space, maintaining the efficiency and productivity of an industrial robot. That is possible thanks to a synergy of control strategies and specific mechanical and electronic solutions, as SEA. The human-robot interaction causes a new problem: the interaction between workers and the collaborative robot. The present thesis deals with robot control and human-robot interaction: a control strategy for guiding collaborative robots using wearable EMG and IMU sensors is presented in this work. The ultimate goal is to make possible to the human operator the control of robot's behavior by the use of sensors. After a first descriptive section over the instrumentation, the design and implementation stages of the control system are presented. The control program is implemented inside the ROS environment. The Myo armband is the sensing device. Two collaborative robots were used to test the system, they have six-degree of freedom and they are called FourByThree and UR10e. The resulted control system can be divided into three different control methods: • Cartesian velocity control: the linear velocity of the arm is used to control the cartesian velocity of the robot. • Teaching by manual guidance and path repetition: the robot is manually moved and a gesture command of the worker is used to save the robot configuration. In that way, the robot learns a path executable by itself. Another gesture command allows the repetition of the last memorized path. • Direction control: the robot is moved along the same direction as the arm movement. The final section of the present work shows the results of the tests on the robots.